OneStream continues to expand its artificial intelligence strategy with the announcement of a new technology layer called Finance Agentic Layer. Presented at the Splash 2026 event, this innovation is designed to enable finance departments to use external AI assistants such as OpenAI ChatGPT, Anthropic Claude and Microsoft Copilot, while preserving the governance and control requirements specific to the finance function.
The publisher’s stated objective is to connect generative AI’s conversational experiences to the governed financial data in OneStream, without compromising security, user permissions or the auditability of information.
According to OneStream, this new approach responds to an evolution in usage patterns observed among companies. More and more organizations are now choosing their generative AI tools at a global level and want to integrate them with their existing financial systems.
The company explains that its Finance Agentic Layer is based on an open protocol that enables different artificial intelligence systems to communicate securely with one another. This architecture is intended, in particular, to allow users to query ChatGPT or Copilot in natural language and obtain financial analyses, dashboards or projections directly powered by OneStream data.
In the examples presented by the publisher, an executive can ask for a summary of financial trends, budget variances or forecasts, and then receive an automated output generated from the platform’s consolidated data.
OneStream nevertheless emphasizes a central point: financial controls remain in force even when users access external AI tools. Each interaction is authenticated according to the access rights defined in OneStream, while administrators retain full visibility into usage and the activities carried out.
This announcement is part of a broader trend in the EPM and FP&A software market, where vendors are accelerating the development of AI agents capable of automating certain analysis, forecasting and financial reporting tasks.
A new “Forecast Agent” was also presented by OneStream. It is intended to enable users to ask questions in natural language about financial forecasts and receive contextualized answers based on the company’s data.
With this development, OneStream is seeking to position its platform as a financial intelligence layer capable of powering different AI ecosystems, while preserving data reliability and the governance standards expected by finance departments.
Source: Onestream
